The redshift men
The Mount Wilson Observatory in California had been built around a telescope with a 60-inch reflecting mirror, which came into operation in 1908. Just ten years later, this was joined on the mountain by the 100-inch Hooker Telescope (named after the benefactor who paid for it), which was to be the most powerful astronomical telescope on Earth for nearly 30 years, until the completion of the famous 200-inch Hale Telescope (named after George Ellery Hale, the astronomer who created both the Mount Wilson and the Mount Palomar observatories), at Mount Palomar, near Los Angeles (not far from Pasadena), in 1947. There were two people who would push the 100-inch to its limits in the 1920s.
The first of those pioneers, Milton Humason, was born in Dodge Center, Minnesota, on 19 August 1891; but his parents moved the family to the West Coast when he was a child. At the age of 14, in 1905, Humason was taken to a summer camp on Mount Wilson (this was about the time the observatory was being established), and fell in love with the mountain. He persuaded his parents to let him take a year out from his education, and got a job at the then new Mount Wilson Hotel (lower down the mountain than the observatory), working as a bellboy and general handyman, and looking after the pack animals which were used in those days to carry goods (and people) up the mountain trails.
Humason never went back to school. Instead, by the end of the decade he became a mule driver, working with pack trains carrying equipment right up to the peak of the mountain, where the 60©-nch reflector (then the best astronomical telescope in the world) had become operational, and work was in progress on the dome and other buildings associated with the planned 100©inch telescope. Every
item of equipment for the observatory, from the telescopes themselves, to lumber and other building supplies, and the food for the construction gangs and the astronomers themselves, went up the mountain this way. This is as good an indication as any of just how much technology has changed since the early 20th century, and just what an achievement the 100©inch was in its day. There was also the minor point that anyone working on the mountain had to keep a careful lookout for mountain lions, which still roamed the peak then.
While working on the mule trains and enjoying the outdoor life, Humason fell in love with Helen Dowd, the daughter of the engineer in charge of the activities on the mountain peak, and the couple were married in 1911, when they were both just 20 years old. The arrival of a baby, William, in the autumn of 1913 persuaded Milton that he ought at last to think about putting down some roots, and for three years he worked as head gardener on an estate in Pasadena (some reports describe this as being “foreman on a ranch”, but even in 1914 Pasadena wasn’t exactly the Wild West; the term “ranch” wasoften used in the same way that we would use the word “farm”).
Three years later, the young couple purchased their own “citrus ranch”; but almost immediately an opportunity came up that Milton and Helen, who had been pining for the mountain, couldn’t resist. Helen’s father told them that one of the janitors at the observatory was about to leave, and suggested that the job might suit young Milton. Even better, with the 100©inch telescope due to become operational in 1918, there was a chance to combine the janitorial duties with the post of “relief night assistant”, helping out the astronomers, if required, on both the big telescopes. The pay was modest — $80 per month — but the post included rent-free accommodation, and free meals while working. And it meant living on the mountain (by all accounts, if he had had any money Humason would have paid themïto let him live on the mountain). He took up the post in November 1917.
Within a year, Humason had learned how to take photographic plates of astronomical objects, using the smaller telescopes on the mountain, and he proved so adept at this arcane art that in 1920 hewas officially appointed to the astronomical staff of the observatory ). There were some mutterings about this promotion of the high school dropout and mule skinner, who just happened to be the son-in-law of the observatory’s chief engineer; but these were soon stilled as Humason’s remarkable ability at obtaining astronomical photographs
His boss, Harlow Shapley, described Humason as “one of the best observers we ever had”; from a distance of nearly 100 years, he looks like the bestobserver on the mountain in the 1920s and 1930s. And this was quite an achievement. The arcane skills involved in getting images of faint astronomical objects in those days began with the actual observations. This meant sitting at the telescope night after night (perhaps every night for a week) keeping it pointing accurately a the object of interest (typically, a galaxy, in Humason’s case) while the light from the object was gathered in and directed to a glass photographic plate (coated with light-sensitive material) at the focus of the telescope. In those pre-computer days, the telescope needed constant human attention to keep tracking perfectly across the sky to compensate for the rotation of the Earth and hold the same celestial object centred in its field of view for hours on end — it did have an automatic tracking system (essentially a clockwork mechanism controlling electric motors) but this had its own little foibles and could not be left unattended. And, of course, the dome had to be open to the sky, so the telescope could see out, and it had to be unheated, because convection currents of air rising past the telescope would blur its field of view. Even in summer, the mountain top can be cold at night (I visited it in May one year, when there was snow on the ground); and the best time to observe, of course is in the depths of winter, when the skies are dark for longest. One other thing — there could be no artificial light inside the dome, apart from a dim red bulb, because that would fog the photographic plates.
Each night, working under these difficult conditions, the same plate would be carefully exposed to the light from the telescope at the start of the observing run, and carefully shut away in a dark container at the end of the night’s observing. Only after a week or so would enough light have been gathered to provide a good image of the object. Then, the observer would have to process the plate, by hand, in the dark (a fragile glass plate, remember), using a variety of chemicals first to develop the picture, and then to fix it as a permanent image on the plate. Use the wrong strength of chemicals, or apply them for the wrong amount of time (or let the plate slip from your grasp), and a week’s work would be ruined. Extreme patience and a calm, unflappable manner were essential requirements for a successful observational astronomer in those days — as it happens, characteristics that are also required of a successful mule driver.
Even though he became the best observer on Mount Wilson, and probably the best in the world, Humason was always diffident about his lack of academic qualifications, and understandably cautious (especially in his early years as an astronomer) about pushing his own ideas forward. The combination of his great skill at astronomical photography and this understandable diffidence led to a bizarre incident, which happened shortly before Shapley left the mountain to take up his post at Harvard. This was early in 1921, the year after Humason had been appointed to the astronomical staff in the most junior capacity, and the year before he received the dizzying promotion to the rank of assistant astronomer. Humason had been given the task (by Shapley) of comparing plates of the Andromeda Nebula, M31, obtained by the new 100-inch telescope on different occasions, to see if there were any differences in the images (this was probably in an attempt by Shapley to find evidence for the kind of rotation that had been claimed by some astronomers for other nebulae). The way this kind of comparison was made (and still is, on some occasions, although computers have largely taken over) was to “blink” the plates in a special kind of viewer. Looking through the eyepiece of this device, you see each plate in turn, repeatedly, with the image bouncing backward and forward between the two. When this is done, any differences in the two images leap out at the human eye.
To Humason’s surprise, when he blinked the plates of the Andromeda Nebula in this way, he thought he could see tiny specks of light that were present on some plates but not on others — as if there were variable stars in the nebula. He carefully took the plate with the best example of this, and marked the positions of the interesting features with little lines drawn in ink on the back of the plate. Then, he took the plate to show Shapley what he had found. Shapley simply ignored Humason’s claim. First, he explained to the most junior astronomer on the mountain just why it was impossible for there to be variable stars in the Andromeda Nebula. Then, he took a clean handkerchief out of his pocket, turned the plate Humason had given him over, and wiped away the identifying ink marks. A few weeks later, on 15 March 1921, he left for Harvard.
Humason said nothing to anyone at the time, for obvious reasons. He had barely got his foot on the first rung of the astronomical ladder, and he owed even that modest position largely to Shapley’s recommendation. But later in his career he told the tale on several occasions; one of the interested listeners was
Allan Sandage, who later had a big part to play in the investigation of the Universe. There are many tantalising “what ifs” hanging around the story. If Shapley had stayed at the Mount Wilson Observatory, might he have had second
thoughts, and discovered the truth about the spiral nebulae? Or would his stubbornness have had an influence on his colleagues there, and held back the discovery of this truth? It is a fruitless game to play, but the moral of the story is clear — you have to accept the observations (or at least, take them seriously enough to look at them in detail), even if they conflict with your cherished theory.
The other pioneer who, together with Humason, reshaped our understanding of the Universe in the 1920s, took that attitude to extremes. Edwin Hubble never really subscribed to any theory about the Universe at all, in spite of the association made today between his name and the theory of the Big Bang. Hubble was an observer, and he reported the observations he made almost entirely without any trappings of theoretical interpretation, leaving that for others to do. He also came from a background of academic achievement that contrasts sharply with the background of Humason, with whom his name will always be linked — although, as we shall see, Hubble always exaggerated his own social status and achievements outside astronomy.
Hubble had been born in Marshfield, Missouri, in 1889. He was one of eight children; their father, a failed lawyer, worked in insurance and travelled widely (as a manager overseeing scattered offices), so as a child Edwin’s adult male role models were his two grandfathers. It was his maternal grandfather, a medical doctor called William James, who, we are told, introduced Edwin to the wonders of astronomy by building his own telescope and allowing the young boy to look through it at the stars as a treat on his eighth birthday.
At the end of 1899, the family moved to Evanston, Illinois, on the shore of Lake Michigan, and in 1901 to the newly incorporated city of Wheaton, just outside Chicago. So it was in Chicago that Edwin Hubble attended both high school and university, making a name as a good athlete (although not quite the all round star that he would later lead people to believe) and as a first class scholar. After studying science and mathematics for two years, and being awarded the two-year Associate in Science Degree, Hubble concentrated on courses in French, the Classics and political economics, aiming for a Rhodes Scholarship, which he duly won. He received his Bachelor’s degree in 1910, then took up the Scholarship at Queen’s College, in Oxford, where he studied law and acquired an exaggerated “Oxford Britishness” in speech and mannerisms that stayed with him for the rest of his life.
Hubble’s father died in 1913, at the early age of 52, a few months before the Rhodes Scholar returned from England. During what must have been a traumatic year that followed, Edwin helped to settle his father’s modest estate and made sure that the family, now living in Louisville, stayed together. In spite of his later claims to the contrary, he never practiced law, but he did work for a year as a high school teacher. His immediate duty by his family done, in 1914 Hubble moved on to the Yerkes Observatory (part of the University of Chicago) as a research student in astronomy (it is perhaps worth mentioning that he was only able to do this because his younger brother, Bill, largely took over the financial responsibility of looking after Hubble’s mother an sisters).
The Yerkes Observatory was the first to have been founded by Hale (who by 1914 had long since moved on to Mount Wilson), using funds provided by the millionaire Charles T. Yerkes, who made his money out of trolley cars. The main instrument there was a 40-inch refracting telescope (one that uses lenses, not mirrors), which was then one of the best astronomical telescopes in the world, and is still the largest refractor ever made (and still in use). Hubble’s main work as a student and research assistant between 1914 and 1917 was to photograph as many of the faint nebulae as possible — by the time he joined the observatory, about 17,000 nebulae had been catalogued, and it was estimated that perhaps ten times more might be visible, in principle, to the 40©inch at Yerkes, the new 60-inch reflector on Mount Wilson, and comparable instruments. But this was still before the distinction between nebulae that are gas clouds and part of the Milky Way and what we would now call galaxies was recognised. Hubble’s first contribution to astronomy was an attempt to classify the nebulae according to their appearance, but although his work was good enough for him to be awarded his PhD in 1917, little came of these efforts for another five years, partly because
of America’s involvement in World War I.
Even before he completed his PhD, Hubble had been offered a post on Mount Wilson by Hale, who had been head-hunting to carry out an increase in staff on the mountain in anticipation of the 100-inch telescope becoming operational, and naturally turned to Yerkes as a source of suitable candidates. In fact, Hubble had wanted to stay at Yerkes, but there were no funds available for him there, so he had little choice but to accept the offer from California. But in April 1917, the United States declared war on Germany, in response to the German policy of unrestricted submarine warfare. Hubble volunteered for the infantry as soon as he had completed the formalities for his PhD, and Hale promised to keep the job at Mount Wilson open for him until he returned from Europe.
Hubble’s own account of his military experiences differs from the official records, although there is no doubt that he achieved the rank of Major. His division, the 86th, reached France only in the last weeks before hostilities ended, and never saw combat. Yet Hubble always said (or implied) that he had been in action and had been wounded by shell fragments, which was why he could not straighten his right elbow properly. He also managed to linger in England, which he loved, for long enough before returning to the United States for an irritated Hale to write urging him to make haste, since the 100-inch was operational and there was plenty of work to do. But it was not until 3 September 1919 that Major Hubble (he liked to use the title even in civilian life) finally joined the staff of the observatory on Mount Wilson, when he was only a couple of months short of his 30th birthday.
Hubble first made his name as an astronomer by developing the ideas from his PhD thesis, and coming up with a classification scheme for galaxies (I shall use the modern term, although Hubble always preferred the word nebulae). One of the important early contributions made by Hubble was the recognition that there are huge numbers of another kind of object, different from the spiral nebulae, which also seemed impossible to explain in terms of phenomena contained within the Milky Way. These are now known as elliptical galaxies. The differences between ellipticals and spirals are completely unimportant for now; all that matters is that in due course it was realised that both kinds of nebula are indeed galaxies in their own right. It is now thought that ellipticals (which range in appearance from spherical to a flattened convex lens shape, like the profile of an American football) are formed by mergers between spirals, explaining (among other things) why the largest galaxies known are ellipticals. But none of this was known to Hubble in the early 1920s. The classification scheme he developed was essentially complete by the summer of 1923, although it wasn’t published until some time later.
While Hubble was gathering his evidence in favour of his classification scheme, and becoming increasingly adept at using the 100-inch, the debate about the nature of the nebulae had continued‹d‹ to flicker. The “island universe” idea, that the nebulae (or at least some of them) are other galaxies like our Milky Way, had been championed by the Swedish astronomer Knut Lundmark in his PhD thesis in 1920, and in 1921 and 1922 he visited both the Lick Observatory and Mount Wilson, obtaining spectra of the spiral known as M33, and convincing himself (but certainly not Shapley) that the speckled, grainy appearance of the nebula meant that it was indeed composed of large numbers of stars. In 1922, three variable stars were identified in the patch of sky covered by M33, but the observations of these very faint objects were not good enough for the nature of these stars to be determined; in 1923, a dozen variables were found in another nebula, NGC 6822, but again the observations were not good enough to identify the nature of these stars immediately (it took a year’s observations before they were eventually identified as the variable stars known as Cepheids, and by then this was no surprise).
The search for Cepheids in nebulae didn’t look too promising in the middle of 1923, when Hubble had completed his work on the classification scheme, but the prospect of finding novae (exploding stars) in the nebula, using the 100-inch, looked much more promising. If ordinary novae could be firmly identified in M31, that would be as good a way as any of establishing the approximate distance to the nebula.
It was with this in mind that Hubble began another observing run with the 100©inch in the autumn of 1923, concentrating on photographing one of the spiral arms in the Andromeda Nebula, M31. Seeing conditions were poor on the night of 4 October, but even so a 40-minute exposure produced a plate with a bright spot, possibly a nova. The next night, a slightly longer exposure confirmed the presence of the nova, and showed two more spots of light — two more suspected novae. Back in his office, Hubble dug out earlier plates showing the same part of M31, going back several years and obtained by various different observers, including Humason and (ironically) Shapley. It was this series of plates which, under close examination, showed that one of the two additional “novae” discovered by Hubble on 5 October was, in fact, a Cepheid variable
with a period of just under 31.5 days. Plugging in the known period-luminosity relationship and distance calibration used by Shapley himself in a survey of the Milky Way Galaxy, this immediately gave Hubble a distance of 300,000 parsecs to the Andromeda Nebula — almost a million light years, and three times the size of what Shapley had considered to be the entire Universe. Since then, partly because of calibration problems that the 1920’s astronomers were unaware of, the estimated distance to M31 has been revised up to about 700 kiloparsecs; but even with the incorrect calibration, Hubble had proved that at least one spiral nebula was indeed an object comparable in size to our Galaxy, and far beyond the Milky Way.
Over the winter months of 1923-24, Hubble found nine novae and another Cepheid in M31, all pointing to the same conclusion. In 1924, he found nine Cepheids in another nebula, NGC 6822, 15 in the spiral M33, and others in other nebulae. Hubble’s place in the history books would have been assured if he had given up astronomy on the spot. But there was another pressing puzzle about the nature of the nebulae, one which cried out for careful study using the best telescope on Earth, the 100-inch. It was a puzzle that had been building up for more than a dozen years, since Hubble was a Rhodes Scholar in Oxford, where he knew nothing of the work being carried out at the great observatories in
the United States.
Pioneering studies by Vesto Slipher, working at the Lowell Observatory, had shown that a few nebulae (all he could study with the equipment he had) show blue- or red- shifts in their spectra, interpreted as indicating that they are moving towards us (blueshift) or away (redshft). Most seemed to have redshifts. In 1926, Slipher was coming to the end of his studies of redshifts, because the equipment he had available, based on a 24-inch refractor, had been pushed to the limit of what it could observe. But there was a hint that fainter, and therefore presumably more distant, “nebulae” had bigger redshifts. Hubble wanted to search for a relationship between redshift and distance, so the first thing he would have to do would be to find distances to as many as possible of the nebulae whose redshifts had been measured by Slipher. But in order to probe deeper into the Universe, as he realised in 1926, he would need redshifts for fainter objects, which could best be obtained by the 100-inch. Hubble himself was deeply involved with the continuing programme of distance measurements, and the 100-inch had never been used for redshift work, involving spectroscopic photographs of very faint objects, before. He needed someone to undertake the taxing task of adapting the telescope to this new work, and then making the measurements themselves.
Humason was the obvious choice, not just because he was a superb observer, but also because of the clear difference in status. Although Hubble knew he had to have help with his latest project, he didn’t want a collaborator of equal status as an astronomer to himself; he wanted an assistant, so that as much as possible (preferably all) of the glory associated with the work would be his. Humason took up the challenge, and in order to test the possibilities he chose for his first attempt at a redshift measurement a nebula which was too faint for its light to have been analysed in this way by Slipher at the Lowell Observatory. After two nights patiently keeping the great telescope tracking the faint nebula, he had a spectrum good enough to show (under a magnifying lens) spectroscopic lines associated with the presence of calcium atoms in the nebula. The lines were shifted toward the red end of the spectrum, by an amount corresponding to a Doppler velocity of some 3,000 kilometers a second, more than twice as large as any redshift measured by Slipher.
The trial run had been a success, but it had also shown Humason how physically demanding it would be to obtain more spectra from faint nebulae. The prospect of spending night after night freezing in his seat at the guidance controls of the telescope, all for the benefit of someone else’s research project, and all to confirm (at least at first) what Slipher had already discovered, did not appeal to him, and he said so in no uncertain terms. He was persuaded to carry on with the task partly by some flattering comments from Hale (who had retired as Director of the Mount Wilson Observatory, on health grounds, but still kept in close touch) and by the promise of a new spectrograph, much more sensitive than the old one, which would enable spectra of even faint nebulae to be obtained in a single night. Humason agreed to carry on. Of course, in the long run the new spectrograph didn’t really ease his burden. If a faint nebula could now be photographed spectroscopically in a single‹d‹ night, then a îveryï faint nebula could be photographed in two or three nights of observation. Astronomers are always pushing their equipment (and in those days, themselves) to the limit. Before long, Humason was hooked on the project, working harder than ever to obtain redshifts for fainter and fainter objects.
But he took things step by step. Showing exemplary caution and patience (he must have been a really good mule driver), in spite of his initial success Humason spent many months bedding the new equipment in, and honing his own skill at the new technique, by re-measuring the redshifts of all 45 nebulae analysed by Slipher. He found the same values of the redshifts that Slipher had found, important confirmation that the results meant something, and that the combination of the 100-inch, the new spectrograph, and Humason himself was ready to take the leap out to higher redshifts.
Meanwhile, Hubble had been making distance measurements (using a variety of techniques) for many of the same nebulae, and had a pretty good idea that the two sets of data showed a linear relationship between redshift and distance — that redshift is proportional to distance, so that if one galaxy has twice as big a redshift as another, it is twice as far away. Indeed, he must have had some idea of this already in 1926, as it had been suggested by the Belgian astronomer Georges Lemaitre in a paper Hubble had seen in draft; but he was extremely cautious about putting this conclusion down in print, and was only pushed into doing so when it looked as if someone else was on the same trail.
The someone else was the Swde Knut Lundmark, who at the end of 1928 made a formal request to the then Director of the Mount Wilson Observatory, Walter Adams, to visit the mountain for the express purpose of measuring the redshifts of faint nebulae. He even asked if Milton Humason might be available to help him in this work. Lundmark was politely rebuffed, and Hubble took the hint, publishing his first short paper on the redshift-distance relationship early in 1929. In that paper (just six pages long, and titled “A Relation Between Distance and Radial Velocity Among Extra-Galactic Nebulae”) Hubble claimed to have accurate distance measurements to just 24 nebulae for which redshifts were widely known at the time, and less accurate distances to another 22. When these measurements were plotted as points on a graph, with distance along the horizontal axis and velocity up the vertical axis, they were scattered rather widely, but with a tendency for higher velocities to be associated with higher redshifts. Hubble drew a straight line through these scattered points, with a slope which set the constant of proportionality in the redshift-distance relation as about 525 kilometers per second per Megaparsec (about 20 per cent less than the value suggested by Lemaitre).
On the evidence of the 1929 paper alone, it is hard to justify choosing this particular slope for the straight line (to be honest, it is hard to justify drawing a straight line at all); but Hubble already knew of at least one galaxy with a much higher redshift and correspondingly greater distance, and it is certain that he chose this particular straight line to make his published results in that 1929 paper line up with the unpublished data for larger redshifts that he was still working on. Why was he so cautious about revealing the new results that were now coming in from a comparison of his own distance work and Humason’s redshifts? Because he wanted to finish the job before publishing a full paper. If other astronomers (such as Lundmark) got wind of just how successful Humason was being in his measurements of very high redshifts, they might get in on the act, and steal some of the thunder from the Mount Wilson team. Sharing the glory with Humason, clearly his junior, might just be acceptable; sharing the glory with someone from a different observatory was not.
Even so, Hubble’s claim of a linear redshift-distance relationship was quickly accepted by the astronomical community, and became known as Hubble’s Law. After all, as we have seen, the idea of some sort of relationship between redshift and distance was very much in the wind, and people were primed to believe it (not least because a linear relationship is the simplest kind, and the easiest to work with). The snag was, that the kind of redshift-distance relation found by Hubble (and the as yet unsung Humason) did not match up with theoretical models that had been developed by Albert Einstein and others.. Eddington commented on this difficulty for the theorists at a meeting of the Royal Astronomical Society, in London, in January 1930 . When Lemaitre read these comments in the published account of the meeting, he wrote to Eddington, enclosing a copy of his own paper and pointing out that the kind of redshift-distance relation found by Hubble could indeed arise naturally in the context of the general theory of relativity. Eddington promptly wrote to Nature, the leading scientific journal of the time, drawing attention to Lemaitre’s work. Almost everyone agreed that Lemaitre had the explanation for the redshift-distance relation discovered by Hubble, and that the Universe as a whole must be physically expanding, getting bigger as time passes.
At the beginning of 1931, Hubble and Humason together published a paper, “The Velocity-Distance Relation Among Extra-Galactic Nebulae”, which at last revealed most of the data which Hubble had been hugging to himself for the past couple of years. With another 50 redshifts, they more than doubled the number in Hubble’s 1929 paper, and pushed the record out to a cluster of galaxies with a redshift corresponding to a velocity of recession of just under 20,000 kilometers per second, at a distance estimated at the time as a little more than a hundred million light years. When the data were plotted as a graph, the straight line, with almost the same slope as in the 1929 paper, was still there; but the scatter in the points along the line was much smaller, and the choice of the slope for the straight line looked much more plausible.
Adapted from my book The Birth of Time